Project Vignette
Phoebe Fei
4/25/2022
Introduction
The naked mole rat (Heterocephalus glaber) is a extraordinarily long-lived mammal, and has been the focus of study for cancer and aging. The maximum lifespan of the naked mole-rat can exceed 30 years, and are resistant to both cancer and tumorigenesis – distinct from its close relative, such as the Guinea pig and the Damaraland mole rat, of similar size, but are susceptible to cancer and do not have a life-span as long (1,2,3).
Naked mole rat genomics haven’t been studied as thorough as other organisms. In 2011, Kim et al. has published a paper on the physiology of naked-mole rat, identifying some candidate genes, for example, SMAD3, that might be responsible for protecting naked-mole rats from cancer (2). However, this conclusion has been challenged by Delsuc et al., as some mutations are shared by the Guniea pigs, and that Kim has failed to account for some overlooked factors in the analysis study (4).
In 2018, Bens et al. conducted a comparative transcriptome study between Naked mole rats and Guinea pigs, and reveals transcriptome patterns in breeder naked mole rats which might contribute to the long life span (1). Their conclusion was that non-breeder naked mole-rats show low levels of differentiation between sexes comparing to non-breeder guinea pig, and after transition into breeders, naked mole rat transcriptomes become remarkably sex-specific. Since this is a study comparing across species, I thought to conduct a comparison within the naked mole-rat population to see if there are differentially expressed genes within the species of interest. Additionally, I was interested in looking at if using different alignment tools produce different results. I downloaded their data on the liver RNAs of naked mole rats, analyzed and compared within the naked mole rat population.
Hypothesis: There are differentially expressed genes between male and female, as well as between breeder and non-breeder naked mole rats. Using STAR or HISAT2 as alignment tool will affect the result.
Results
FASTQ Quality
First, the quality of the fastq sequences obtained are accessed to decide if data processing is needed before alignment. As shown in the figures, the quality, GC contents and duplication are acceptable to continue to alignment without adapter trimming. In addition, in Checkpoint1 I have performed a test sample with trimgalore, and compared the FASTQC between trimmed and untrimmed. The results showed that no much difference is captured. See checkpoint1 html file.
Sequence Counts
GC Contents
### Duplication Level
Alignment Quality
I have performed both STAR and HISAT2 alignments on the same set of RNA data, since comparing different tools might account for some systematic error. Both methods return nice alignment scores of more than 90%.
STAR Alignment Score Plot
STAR alignment score plot shows that for all samples, most of sequences are uniquely mapped, and we have obtained a good amount of counts to proceed to analysis.
HISAT2 Alignment Score Plot
HISAT2 alignment score plot shows that for all samples, most of sequences are uniquely mapped, and we have obtained a good amount of counts to proceed to analysis.
HISAT 2: Alignment Scores 
Read Distribution Plot
Then, the read distribution of both aligned results are accessed. For STAR, most of the reads obtained are exons, which is expected to proceed to analysis. UTR regions can be decreased in the feature counts step.
STAR Read Distribution 
Similar results
are shown for HISAT2 alignement as well.
HISAT2 Read Distribution 
I have
concluded that the quality of alignment is ready to proceed to analyzing
the differential expression.
Gene Body Coverage Plot
Both STAR and HISAT2 did similarly on gene body coverage. The coverage was uniform with sight drop on the 5’ and 3’ ends.
STAR Gene Coverage
HISAT2 Gene Coverage
ReadCount Summary
Female Sample Feature Count Summary
Below is a summary of the read counts data for female samples of both STAR and HISAT2. Because STAR does not show unmapped reads, it explains who HISAT2 has fewer unassigned reads comparing to STAR, while HISAT2 also has unmapped reads.
For the male summary plot, the result was similar.
Male Sample Feature Count Summary
DESeq Quality
PCA
For the DESeq analysis, I have included sex and breed as design conditions, and below are the PCA plots for each features with each method, respectively.
HISAT2
Breed
Sex
HISAT2
Breed
Sex
They indeed account for some variations within the population, for both
alignment measures.
Differential Expression Analysis
Heatmap
As an overview, the scaled heatmap for breed status test and sex test, and for both aligners are shown below.
Sex:STAR
Sex:HISAT2
Breed:STAR
Breed:HISAT2
Specific Genes with Most Test Significance
Shown below are the top 10 differential expressed genes when tested against sex and breed status, for HISAT2 and STAR respectively. The top 100 genes are written into .csv files and saved to the git repository.
They are:
Results_test_breed_HISAT2_top100.csv
Results_test_breed_STAR_top100.csv
Results_test_sex_HISAT2_top100.csv
Results_test_sex_STAR_top100.csv
Top 10 Differentially Regulated Genes (Sex: Female vs. Male)
STAR
HISAT2 
Both aligners give similar results, but there are still some differences on the ranking and adjusted p-values.
Top 10 Differentially Regulated Genes (Breed: Non-breeder vs. Breeder)
STAR
HISAT2
I have then identified some candidate genes (with gene names) to
specifically look at their expression between groups.
Sex: PNPLA4
PNPLA4 encodes for an enzyme that has both triacylglycerol lipase and transacylase activities and may be involved in adipocyte triglyceride homeostasis.
STAR
HISAT2
Sex: RAB9A
RAB9A enables GDP binding activity; GTP binding activity; and GTPase activity. Involved in negative regulation by host of symbiont catalytic activity; positive regulation of exocytosis; and regulation of protein localization.
STAR
HISAT2
Breed: MYH7
MYH7 encodes the beta (or slow) heavy chain subunit of cardiac myosin.
STAR
HISAT2
We see that sex is indeed a covariate that should be accounted for for these genes shown above. However, such difference is not especially shown for the testing for breed status. It seems that a specific subset shows significant difference in expression for MYH7.
Volcano Plots
At last, I plotted out all the up and down regulated genes when performing the 2 tests, using STAR and HISAT2 as alignment tool, respectively.
STAR: Volcano for female vs. male
HISAT2: Volcano for female vs. male
####
STAR: Volcano for female vs. male
HISAT2: Volcano for female vs. male
Conclusion
STAR vs. HISAT2
In this analysis, both tools performed similarly on the alignment, although because STAR does not output unmapped results, and other algorithm differences, the output gene ranking by adjusted p-value is slightly different.
Sex as a Covariate
I conclude that there are several differentially expressed genes between the male ans female naked mole rat population. When comparing across species, it is suggested that this factor is considered.
Non-breeder vs. breeder
I conclude that on liver expression, there are no significant differential expression genes between the non-breeder and breeder population. Although, directly graphing the genes showed that there are variabilities within the breeder/non-breeder sample. Smore samples would be helpful to determine if the few outliers have other confounding factors hidden.
Limit of Data
This set of data is probably the one and only set available for naked mole rat expression that fulfills the number of biological and technical replicates for our analysis, so we just have to trust the technical process of it – there is no cross validation. Also, for their close relative, like guinea pig and darama mole rat, there is even less datasets that have enough replicate. Also, for the reference, to keep gene annotations consistent, I had to use female referencing index for all - this will be challenging when applying on other organs.
Methods
File Download and Preparation
1. File Download
The reference genomes for naked mole rats are found in Ensembl. Only female reference genes are used, since the male genomes have unconventional naming for gene_id. Liver differential expression RNA analysis will be conducted between male and female, as well as breeder and non-breeder samples.
Female gtf are located here
The data is accessible from NCBI SRA/GEO dataset under project number PRJNA386015. The link to the SRA run selector is here. There are 12 male samples and 11 female samples.
a. Copy all NMR liver files to the project
#download the metadata file (liver for all NMR) and put it into my directory
scp "C:/Users/tof4003/Desktop/WCM/W2022/CMPB 5004/Project/SraRunTable_liver.txt" tof4003@aphrodite.med.cornell.edu:/athena/angsd/scratch/tof4003/project
head -5 SraRunTable_liver.txtLooking into Accession Table
Run,Age,Assay Type,AvgSpotLen,Bases,BioProject,BioSample,Bytes,Center Name,Consent,DATASTORE filetype,DATASTORE provider,DATASTORE region,Experiment,GEO_Accession (exp),Instrument,LibraryLayout,LibrarySelection,LibrarySource,Organism,Platform,ReleaseDate,reproductive_status,Sample Name,sex,source_name,SRA Study,tissue,strain
SRR5517432,3.7 years,RNA-Seq,50,1143144100,PRJNA385850,SAMN06917366,753974373,GEO,public,"fastq,sra","gs,ncbi,s3","gs.US,ncbi.public,s3.us-east-1",SRX2790801,GSM2609133,Illumina HiSeq 2500,SINGLE,cDNA,TRANSCRIPTOMIC,Heterocephalus glaber,ILLUMINA,2018-06-14T00:00:00Z,breeder,GSM2609133,female,liver,SRP106650,liver,
SRR5517435,3.6 years,RNA-Seq,50,1263344450,PRJNA385850,SAMN06917363,850450852,GEO,public,"fastq,sra","gs,ncbi,s3","gs.US,ncbi.public,s3.us-east-1",SRX2790804,GSM2609136,Illumina HiSeq 2500,SINGLE,cDNA,TRANSCRIPTOMIC,Heterocephalus glaber,ILLUMINA,2018-06-14T00:00:00Z,breeder,GSM2609136,female,liver,SRP106650,liver,
SRR5517437,3.7 years,RNA-Seq,50,1109269000,PRJNA385850,SAMN06917361,731997749,GEO,public,"fastq,sra","gs,ncbi,s3","gs.US,ncbi.public,s3.us-east-1",SRX2790806,GSM2609138,Illumina HiSeq 2500,SINGLE,cDNA,TRANSCRIPTOMIC,Heterocephalus glaber,ILLUMINA,2018-06-14T00:00:00Z,breeder,GSM2609138,female,liver,SRP106650,liver,
SRR5517441,2.4 years,RNA-Seq,51,1253621004,PRJNA385850,SAMN06917357,846076900,GEO,public,"fastq,sra","gs,ncbi,s3","gs.US,ncbi.public,s3.us-east-1",SRX2790810,GSM2609142,Illumina HiSeq 2500,SINGLE,cDNA,TRANSCRIPTOMIC,Heterocephalus glaber,ILLUMINA,2018-06-14T00:00:00Z,non-breeder,GSM2609142,female,liver,SRP106650,liver,
The run is field 1, reproduction status is field 28, sex is field 30
#download data
awk -F "," '{ if($30 == "male") print $1","$28}' SraRunTable_liver.txt > liver_male.txt
awk -F "," '{ if($30 == "female") print $1","$28}' SraRunTable_liver.txt > liver_female.txt
mkdir -p lvr_male
mkdir -p lvr_female
#had to sbatch because this process took a long time
nano data_acc.sh
#! /bin/bash -l
#SBATCH --partition=angsd_class
#SBATCH --account=angsd_class
#SBATCH --nodes=1
#SBATCH --ntasks=1
#SBATCH --job-name=SRA_run
#SBATCH --time=15:00:00 #HH/MM/SS
#SBATCH --mem=10G
#load SRA toolkit
spack load sra-toolkit@2.8.2-1%gcc@6.3.0
cd /athena/angsd/scratch/tof4003/project
#clear the sra directory
rm -r /home/tof4003/ncbi
while read line; do
echo $line
acc=$(cut -d "," -f 1 <<< "$line")
breed=$(cut -d "," -f 2 <<< "$line")
prefetch $acc
#change to fastq format
fastq-dump /home/tof4003/ncbi/public/sra/$acc.sra -O /athena/angsd/scratch/tof4003/project/lvr_male
#rename the file
mv /athena/angsd/scratch/tof4003/project/lvr_male/$acc.fastq /athena/angsd/scratch/tof4003/project/lvr_male/$acc.m_$breed.fastq
done < /athena/angsd/scratch/tof4003/project/liver_male.txt
rm -r /home/tof4003/ncbi
while read line; do
echo $line
acc=$(cut -d "," -f 1 <<< "$line")
breed=$(cut -d "," -f 2 <<< "$line")
prefetch $acc
#change to fastq format
fastq-dump /home/tof4003/ncbi/public/sra/$acc.sra -O /athena/angsd/scratch/tof4003/project/lvr_female
#rename the file
mv /athena/angsd/scratch/tof4003/project/lvr_female/$acc.fastq /athena/angsd/scratch/tof4003/project/lvr_female/$acc.f_$breed.fastq
done < /athena/angsd/scratch/tof4003/project/liver_female.txt
sbatch data_acc.sh
#cleaning the temp files
rm -r /home/tof4003/ncbi
#download reference
mkdir -p NMR_ref
nano ref_d.sh
#! /bin/bash -l
#SBATCH --partition=angsd_class
#SBATCH --account=angsd_class
#SBATCH --nodes=1
#SBATCH --ntasks=1
#SBATCH --job-name=SRA_run
#SBATCH --time=15:00:00 #HH/MM/SS
#SBATCH --mem=10G
wget -O NMR_ref/femaleNMR.gtf.gz http://ftp.ensembl.org/pub/current_gtf/heterocephalus_glaber_female/Heterocephalus_glaber_female.HetGla_female_1.0.105.gtf.gz
#need to also download the reference fasta dna file
wget -O NMR_ref/NMR_female_dna.fa.gz http://ftp.ensembl.org/pub/current_fasta/heterocephalus_glaber_female/dna/Heterocephalus_glaber_female.HetGla_female_1.0.dna.toplevel.fa.gz
#unzip .gtf and .fa files
zcat /athena/angsd/scratch/tof4003/project/NMR_ref/femaleNMR.gtf.gz > /athena/angsd/scratch/tof4003/project/NMR_ref/femaleNMR.gtf
zcat /athena/angsd/scratch/tof4003/project/NMR_ref/NMR_female_dna.fa.gz > /athena/angsd/scratch/tof4003/project/NMR_ref/NMR_ref_female.fa
sbatch ref_d.sh
head -10 NMR_ref/femaleNMR.gtfFeatures in the .gtf file
#!genome-build HetGla_female_1.0
#!genome-version HetGla_female_1.0
#!genome-date 2012-02
#!genome-build-accession GCA_000247695.1
#!genebuild-last-updated 2017-07
JH602043.1 ensembl gene 4190633 4194963 . - . gene_id "ENSHGLG00000010267"; gene_version "1"; gene_name "ZMYND10"; gene_source "ensembl"; gene_biotype "protein_coding";
JH602043.1 ensembl transcript 4190633 4194963 . - . gene_id "ENSHGLG00000010267"; gene_version "1"; transcript_id "ENSHGLT00000014491"; transcript_version "1"; gene_name "ZMYND10"; gene_source "ensembl"; gene_biotype "protein_coding"; transcript_name "ZMYND10-201"; transcript_source "ensembl"; transcript_biotype "protein_coding";
JH602043.1 ensembl exon 4194756 4194963 . - . gene_id "ENSHGLG00000010267"; gene_version "1"; transcript_id "ENSHGLT00000014491"; transcript_version "1"; exon_number "1"; gene_name "ZMYND10"; gene_source "ensembl"; gene_biotype "protein_coding"; transcript_name "ZMYND10-201"; transcript_source "ensembl"; transcript_biotype "protein_coding"; exon_id "ENSHGLE00000106764"; exon_version "1";
JH602043.1 ensembl CDS 4194756 4194847 . - 0 gene_id "ENSHGLG00000010267"; gene_version "1"; transcript_id "ENSHGLT00000014491"; transcript_version "1"; exon_number "1"; gene_name "ZMYND10"; gene_source "ensembl"; gene_biotype "protein_coding"; transcript_name "ZMYND10-201"; transcript_source "ensembl"; transcript_biotype "protein_coding"; protein_id "ENSHGLP00000014347"; protein_version "1";
JH602043.1 ensembl start_codon 4194845 4194847 . - 0 gene_id "ENSHGLG00000010267"; gene_version "1"; transcript_id "ENSHGLT00000014491"; transcript_version "1"; exon_number "1"; gene_name "ZMYND10"; gene_source "ensembl"; gene_biotype "protein_coding"; transcript_name "ZMYND10-201"; transcript_source "ensembl"; transcript_biotype "protein_coding";It will be useful to keep the gene_name and gene_biotype in the analysis.
b. creating Index
Create STAR index for NMR female, with editing the resource request to speed up as suggested.
mkdir -p star_liver
nano star_liver/star_female_ref.sh
#! /bin/bash -l
#SBATCH --partition=angsd_class
#SBATCH --account=angsd_class
#SBATCH --nodes=1
#SBATCH #SBATCH --cpus-per-task=4 #try multiple threads with multiple cpu
#SBATCH --job-name=REf_gen_STAR
#SBATCH --time=48:00:00 #HH/MM/SS
#SBATCH --mem=100G
cd /athena/angsd/scratch/tof4003/project/NMR_ref
mkdir -p STAR_ref_female
#load STAR
spack load star@2.7.0e
STAR --runMode genomeGenerate \
--runThreadN 4 \
--genomeDir STAR_ref_female \
--genomeFastaFiles NMR_ref_female.fa \
--sjdbGTFfile femaleNMR.gtf \
--limitGenomeGenerateRAM 100326568150 \
--sjdbOverhang 99
sbatch star_liver/star_female_ref.shc. Checking Quality of fastqc files
Run fastqc, then Multiqc on the fastq files.
cd /athena/angsd/scratch/tof4003/project
#freeing some memories
nano gz.sh
#! /bin/bash -l
#SBATCH --partition=angsd_class
#SBATCH --account=angsd_class
#SBATCH --nodes=1
#SBATCH --ntasks=1
#SBATCH --job-name=zip_files
#SBATCH --time=15:00:00 #HH/MM/SS
#SBATCH --mem=8G
cd /athena/angsd/scratch/tof4003/project
gzip lvr_male/*.fastq lvr_male/*.fastq.gz
rm lvr_male/*.fastq
gzip lvr_female/*.fastq lvr_female/*fastq.gz
rm lvr_female/*.fastq
sbatch gz.sh
mkdir -p fastq_qc
cd fastq_qc
nano fastqc.sh
#! /bin/bash -l
#SBATCH --partition=angsd_class
#SBATCH --account=angsd_class
#SBATCH --nodes=1
#SBATCH --job-name=fastqc
#SBATCH --time=12:00:00 #HH/MM/SS
#SBATCH --mem=8G
spack load fastqc
cd /athena/angsd/scratch/tof4003/project/lvr_male
for file in *.fastq.gz ; do
echo $file
fastqc $file -o /athena/angsd/scratch/tof4003/project/fastq_qc
done
cd /athena/angsd/scratch/tof4003/project/lvr_female
for file in *.fastq.gz ; do
echo $file
fastqc $file -o /athena/angsd/scratch/tof4003/project/fastq_qc
done
#then, run a multiqc
cd /athena/angsd/scratch/tof4003/project/fastq_qc
spack load -r py-multiqc
multiqc .
sbatch fastqc.sh
ls fastq_qc
-exit-
#file download in cmd
scp tof4003@aphrodite.med.cornell.edu:/athena/angsd/scratch/tof4003/project/fastq_qc/multiqc_report.html "/Users/phoebefei/Desktop/WCM/W2022/CMPB 5004/Project/ANGSD Project tof4003/QC htmls/fastqc_report.html"
Alignment
For the purpose of the question, both STAR and Hisat2 were run to access if the expression results differ. Performance-wise, STAR has taken more time and memory. Hisat2 has had generally stable performance.
1. Hisat2
mkdir -p liver_hisat2
cd liver_hisat2
nano hisat2_script.sh
#! /bin/bash -l
#SBATCH --partition=angsd_class
#SBATCH --account=angsd_class
#SBATCH --nodes=1
#SBATCH --cpus-per-task=4 #try multiple threads with multiple cpu
#SBATCH --job-name=Hisat2_tof4003
#SBATCH --time=48:00:00 #HH/MM/SS
#SBATCH --mem=60G
#load modules
spack load hisat2
spack load samtools@1.9% gcc@6.3.0
cd /athena/angsd/scratch/tof4003/project/NMR_ref
#Index the genome
mkdir -p hisat2_index_NMR_female
cd hisat2_index_NMR_female
hisat2-build -p 16 /athena/angsd/scratch/tof4003/project/NMR_ref/NMR_ref_female.fa HISAT2_NMR_female
cd /athena/angsd/scratch/tof4003/project/lvr_male
#run hisat2 alignment: reads are in fastq files, threads = 2; files are not pair-end reads
mkdir -p /athena/angsd/scratch/tof4003/project/liver_hisat2/out_male
for file in *.fastq.gz; do
echo "Begin $file"
fv=$(cut -d "." -f 1 <<< "$file")
suf=$(cut -d "." -f 2 <<< "$file")
hisat2 -q -p 2 -x /athena/angsd/scratch/tof4003/project/NMR_ref/hisat2_index_NMR_female/HISAT2_NMR_female --summary-file /athena/angsd/scratch/tof4003/project/liver_hisat2/$fv.$suf.summary.txt -U $file -S /athena/angsd/scratch/tof4003/project/liver_hisat2/out_male/$fv.$suf.hisat2.sam
#use samtools to covert to bam, sort& index
#sort & covert to bam
samtools view -b /athena/angsd/scratch/tof4003/project/liver_hisat2/out_male/$fv.$suf.hisat2.sam | samtools sort -o /athena/angsd/scratch/tof4003/project/liver_hisat2/out_male/$fv.$suf.hisat2.sorted.bam
#index
samtools index /athena/angsd/scratch/tof4003/project/liver_hisat2/out_male/$fv.$suf.hisat2.sorted.bam
echo "Hisat2 for $fv.$suf done"
#remove .sam files to free up space
rm /athena/angsd/scratch/tof4003/project/liver_hisat2/out_male/$fv.$suf.hisat2.sam
done
#females
cd /athena/angsd/scratch/tof4003/project/lvr_female
mkdir -p /athena/angsd/scratch/tof4003/project/liver_hisat2/out_female
for file in *.fastq.gz; do
echo "Begin $file"
fv=$(cut -d "." -f 1 <<< "$file")
suf=$(cut -d "." -f 2 <<< "$file")
hisat2 -q -p 2 -x /athena/angsd/scratch/tof4003/project/NMR_ref/hisat2_index_NMR_female/HISAT2_NMR_female --summary-file /athena/angsd/scratch/tof4003/project/liver_hisat2/$fv.$suf.summary.txt -U $file -S /athena/angsd/scratch/tof4003/project/liver_hisat2/out_female/$fv.$suf.hisat2.sam
#use samtools to covert to bam, sort& index
#sort & covert to bam
samtools view -b /athena/angsd/scratch/tof4003/project/liver_hisat2/out_female/$fv.$suf.hisat2.sam | samtools sort -o /athena/angsd/scratch/tof4003/project/liver_hisat2/out_female/$fv.$suf.hisat2.sorted.bam
#index
samtools index /athena/angsd/scratch/tof4003/project/liver_hisat2/out_female/$fv.$suf.hisat2.sorted.bam
echo "Hisat2 for $fv.$suf done"
#remove .sam files to free up space
rm /athena/angsd/scratch/tof4003/project/liver_hisat2/out_female/$fv.$suf.hisat2.sam
done
sbatch hisat2_script.sh
2. STAR
cd /athena/angsd/scratch/tof4003/project/star_liver
nano star_align.sh
#star actually gives core-dump issue for male genomes if not enough core is reserved.
#! /bin/bash -l
#SBATCH --partition=angsd_class
#SBATCH --account=angsd_class
#SBATCH --nodes=1
#SBATCH --cpus-per-task=8
#SBATCH --job-name=STAR_align
#SBATCH --time=40:00:00 #HH/MM/SS
#SBATCH --mem=90G
#load modules
spack load star@2.7.0e
spack load samtools@1.9% gcc@6.3.0
cd /athena/angsd/scratch/tof4003/project/star_liver
#running STAR
for file in /athena/angsd/scratch/tof4003/project/lvr_male/*.fastq.gz; do
echo "Begin $file"
fv=$(cut -d "/" -f 8 <<< "$file" | cut -d "." -f 1 )
suf=$(cut -d "/" -f 8 <<< "$file" | cut -d "." -f 2)
STAR --runMode alignReads \
--runThreadN 8 \
--genomeDir /athena/angsd/scratch/tof4003/project/NMR_ref/STAR_ref_female \
--readFilesCommand zcat \
--readFilesIn $file \
--outFileNamePrefix $fv.$suf.star. \
--outSAMtype BAM SortedByCoordinate
#use samtools to index
#index
samtools index $fv.$suf.star.Aligned.sortedByCoord.out.bam
echo "star $fv.$suf done"
done
#females
for file in /athena/angsd/scratch/tof4003/project/lvr_female/*.fastq.gz; do
echo "Begin $file"
fv=$(cut -d "/" -f 8 <<< "$file" | cut -d "." -f 1 )
suf=$(cut -d "/" -f 8 <<< "$file" | cut -d "." -f 2)
STAR --runMode alignReads \
--runThreadN 4 \
--genomeDir /athena/angsd/scratch/tof4003/project/NMR_ref/STAR_ref_female \
--readFilesCommand zcat \
--readFilesIn $file \
--outFileNamePrefix $fv.$suf.star. \
--outSAMtype BAM SortedByCoordinate
#use samtools to index
#index
samtools index $fv.$suf.star.Aligned.sortedByCoord.out.bam
echo "star $fv.$suf done"
done
sbatch star_align.shAlignement Quality Check
1. clearing the files
cd /athena/angsd/scratch/tof4003/project
#separate out the female and female files for star
mkdir -p star_liver/female_out
mkdir -p star_liver/male_out
mv star_liver/*f_* star_liver/female_out
mv star_liver/*m_* star_liver/male_out2. Multiqc on the Output Reports
cd /athena/angsd/scratch/tof4003/project
mkdir -p alignment_repo
cd alignment_repo
mkdir -p star
mkdir -p hisat2
#copy the log files to directories
cp /athena/angsd/scratch/tof4003/project/star_liver/male_out/*.Log.final.out star
cp /athena/angsd/scratch/tof4003/project/star_liver/female_out/*.Log.final.out star
cp /athena/angsd/scratch/tof4003/project/liver_hisat2/*.txt hisat2
nano align_multiqc.sh
#! /bin/bash -l
#SBATCH --partition=angsd_class
#SBATCH --account=angsd_class
#SBATCH --nodes=1
#SBATCH --job-name=bam_fastqc
#SBATCH --time=12:00:00 #HH/MM/SS
#SBATCH --mem=14G
cd /athena/angsd/scratch/tof4003/project/alignment_repo
#run multiqc
spack load -r py-multiqc
cd star
multiqc .
cd /athena/angsd/scratch/tof4003/project/alignment_repo/hisat2
multiqc .
sbatch align_multiqc.sh
#file download in cmd
scp tof4003@aphrodite.med.cornell.edu:/athena/angsd/scratch/tof4003/project/alignment_repo/hisat2/multiqc_report.html "/Users/phoebefei/Desktop/WCM/W2022/CMPB 5004/Project/ANGSD Project tof4003/QC htmls/hisat2_qc.html"
scp tof4003@aphrodite.med.cornell.edu:/athena/angsd/scratch/tof4003/project/alignment_repo/star/multiqc_report.html "/Users/phoebefei/Desktop/WCM/W2022/CMPB 5004/Project/ANGSD Project tof4003/QC htmls/star_qc.html"3. RSeQC on bam file
a. Distribution
Download the .bed file from UCSC Table Browser
.bed field input in UCSC 
Import and edit the .bed file
cd /athena/angsd/scratch/tof4003/project
mkdir -p RSeQC
--exit--
scp "/Users/phoebefei/Desktop/WCM/W2022/CMPB 5004/Project/ANGSD Project tof4003/NMR_female_ref.gz" tof4003@aphrodite.med.cornell.edu:/athena/angsd/scratch/tof4003/project/RSeQC/
cd /athena/angsd/scratch/tof4003/project/RSeQC
gunzip NMR_female_ref.gz
mv NMR_female_ref NMR_ref.bed
#my alignments all have a .1 after the chromosome name
awk -F "\t" '{OFS = "\t"}{$1=$1".1"; print $0}' < NMR_ref.bed > NMR_ref.1.bedRunning distribution on .bam files, STAR and HISAT2 Separate
mkdir -p distribution
mkdir -p distribution/star
mkdir -p distribution/hisat2
nano dis.sh
#! /bin/bash -l
#SBATCH --partition=angsd_class
#SBATCH --account=angsd_class
#SBATCH --nodes=1
#SBATCH --job-name=distribution
#SBATCH --time=12:00:00 #HH/MM/SS
#SBATCH --mem=8G
spack load -r py-rseqc@2.6.4
#star
cd /athena/angsd/scratch/tof4003/project/star_liver/male_out
for file in *.bam; do
echo $file
fv=$(awk -F "." '{print $1"."$2}' <<< $file)
read_distribution.py -i ${file} \
-r /athena/angsd/scratch/tof4003/project/RSeQC/NMR_ref.1.bed > \
/athena/angsd/scratch/tof4003/project/RSeQC/distribution/star/${fv}_rseqc_distribution.out
done
cd /athena/angsd/scratch/tof4003/project/star_liver/female_out
for file in *.bam; do
echo $file
fv=$(awk -F "." '{print $1"."$2}' <<< $file)
read_distribution.py -i ${file} \
-r /athena/angsd/scratch/tof4003/project/RSeQC/NMR_ref.1.bed > \
/athena/angsd/scratch/tof4003/project/RSeQC/distribution/star/${fv}_rseqc_distribution.out
done
#hisat2
cd /athena/angsd/scratch/tof4003/project/liver_hisat2/out_male
for file in *.bam; do
echo $file
fv=$(awk -F "." '{print $1"."$2}' <<< $file)
read_distribution.py -i ${file} \
-r /athena/angsd/scratch/tof4003/project/RSeQC/NMR_ref.1.bed > \
/athena/angsd/scratch/tof4003/project/RSeQC/distribution/hisat2/${fv}_rseqc_distribution.out
done
cd /athena/angsd/scratch/tof4003/project/liver_hisat2/out_female
for file in *.bam; do
echo $file
fv=$(awk -F "." '{print $1"."$2}' <<< $file)
read_distribution.py -i ${file} \
-r /athena/angsd/scratch/tof4003/project/RSeQC/NMR_ref.1.bed > \
/athena/angsd/scratch/tof4003/project/RSeQC/distribution/hisat2/${fv}_rseqc_distribution.out
done
#run multiqc, note that it is best to run multiqc outside the .sh file or as a different job, as sometimes the 2 packages (py-rseqc@2.6.4 & py-multiqc) interfere permission
spack load -r py-multiqc
cd /athena/angsd/scratch/tof4003/project/RSeQC/distribution/star
multiqc .
cd /athena/angsd/scratch/tof4003/project/RSeQC/distribution/hisat2
multiqc .
sbatch dis.sh
b. Coverage
Running coverage on .bam files, STAR and HISAT2 Separate. This step takes amazingly long time
cd /athena/angsd/scratch/tof4003/project/RSeQC
mkdir -p coverage
mkdir -p coverage/star
mkdir -p coverage/hisat2
nano coverage.sh
#! /bin/bash -l
#SBATCH --partition=angsd_class
#SBATCH --account=angsd_class
#SBATCH --nodes=1
#SBATCH --job-name=coverage
#SBATCH --time=50:00:00 #HH/MM/SS
#SBATCH --mem=8G
spack load -r py-rseqc@2.6.4
#star
cd /athena/angsd/scratch/tof4003/project/star_liver/male_out
for file in *.bam; do
echo $file
fv=$(awk -F "." '{print $1"."$2}' <<< $file)
geneBody_coverage.py -i $file \
-r /athena/angsd/scratch/tof4003/project/RSeQC/NMR_ref.1.bed \
-o /athena/angsd/scratch/tof4003/project/RSeQC/coverage/star/${fv}_rseqc_coverage.out
done
cd /athena/angsd/scratch/tof4003/project/star_liver/female_out
for file in *.bam; do
echo $file
fv=$(awk -F "." '{print $1"."$2}' <<< $file)
geneBody_coverage.py -i $file \
-r /athena/angsd/scratch/tof4003/project/RSeQC/NMR_ref.1.bed \
-o /athena/angsd/scratch/tof4003/project/RSeQC/coverage/star/${fv}_rseqc_coverage.out
done
#hisat2
cd /athena/angsd/scratch/tof4003/project/liver_hisat2/out_male
for file in *.bam; do
echo $file
fv=$(awk -F "." '{print $1"."$2}' <<< $file)
geneBody_coverage.py -i $file \
-r /athena/angsd/scratch/tof4003/project/RSeQC/NMR_ref.1.bed\
-o /athena/angsd/scratch/tof4003/project/RSeQC/coverage/hisat2/${fv}_rseqc_coverage.out
done
cd /athena/angsd/scratch/tof4003/project/liver_hisat2/out_female
for file in *.bam; do
echo $file
fv=$(awk -F "." '{print $1"."$2}' <<< $file)
geneBody_coverage.py -i $file \
-r /athena/angsd/scratch/tof4003/project/RSeQC/NMR_ref.1.bed \
-o /athena/angsd/scratch/tof4003/project/RSeQC/coverage/hisat2/${fv}_rseqc_coverage.out
done
#run multiqc, note that it is best to run multiqc outside the .sh file or as a different job, as sometimes the 2 packages (py-rseqc@2.6.4 & py-multiqc) interfere permission
spack load -r py-multiqc
cd /athena/angsd/scratch/tof4003/project/RSeQC/coverage/star
multiqc .
cd /athena/angsd/scratch/tof4003/project/RSeQC/coverage/hisat2
multiqc .
sbatch coverage.shfile download
scp tof4003@aphrodite.med.cornell.edu:/athena/angsd/scratch/tof4003/project/RSeQC/distribution/hisat2/multiqc_report.html "/Users/phoebefei/Desktop/WCM/W2022/CMPB 5004/Project/ANGSD Project tof4003/QC htmls/ResQC_hisat2_distribution.html"
scp tof4003@aphrodite.med.cornell.edu:/athena/angsd/scratch/tof4003/project/RSeQC/distribution/star/multiqc_report.html "/Users/phoebefei/Desktop/WCM/W2022/CMPB 5004/Project/ANGSD Project tof4003/QC htmls/ResQC_star_distribution.html"
scp tof4003@aphrodite.med.cornell.edu:/athena/angsd/scratch/tof4003/project/RSeQC/coverage/hisat2/multiqc_report.html "/Users/phoebefei/Desktop/WCM/W2022/CMPB 5004/Project/ANGSD Project tof4003/QC htmls/ResQC_hisat2_coverage.html"
scp tof4003@aphrodite.med.cornell.edu:/athena/angsd/scratch/tof4003/project/RSeQC/coverage/star/multiqc_report.html "/Users/phoebefei/Desktop/WCM/W2022/CMPB 5004/Project/ANGSD Project tof4003/QC htmls/ResQC_star_coverage.html"Feature Counts
Feature counts are generated for STAR and HISAT2 separately. Here are the flags explained:
-Not including the “-O” (reads can be assigned to more than 1 matched exon, so if reads matched more than one exon, all of them will be counted). As discussed in the homework, the ambiguous reads are most likely UTR’s so will be disgarded -Deleting the -f flag so that only mapped genes are included -Include –extraAttributes for gene_biotype and gene_name
cd /athena/angsd/scratch/tof4003/project
#will generate count tables for star and hisat2 separately
mkdir -p featureCounts
cd featureCounts
nano featCounts.sh
#! /bin/bash -l
#SBATCH --partition=angsd_class
#SBATCH --account=angsd_class
#SBATCH --nodes=1
#SBATCH --job-name=feat_counts
#SBATCH --time=12:00:00 #HH/MM/SS
#SBATCH --mem=15G
cd /athena/angsd/scratch/tof4003/project
spack load subread
featureCounts -a NMR_ref/femaleNMR.gtf \
-o featureCounts/star_featCounts_maleNMR.txt \
-t exon \
-g gene_id \
--extraAttributes gene_name,gene_biotype \
-G NMR_ref/NMR_ref_female.fa \
star_liver/male_out/*.bam
featureCounts -a NMR_ref/femaleNMR.gtf \
-o featureCounts/star_featCounts_femaleNMR.txt \
-t exon \
-g gene_id \
--extraAttributes gene_name,gene_biotype \
-G NMR_ref/NMR_ref_female.fa \
star_liver/female_out/*.bam
featureCounts -a NMR_ref/femaleNMR.gtf \
-o featureCounts/hisat2_featCounts_maleNMR.txt \
-t exon \
-g gene_id \
--extraAttributes gene_name,gene_biotype \
-G NMR_ref/NMR_ref_female.fa \
liver_hisat2/out_male/*.bam
featureCounts -a NMR_ref/femaleNMR.gtf \
-o featureCounts/hisat2_featCounts_femaleNMR.txt \
-t exon \
-g gene_id \
--extraAttributes gene_name,gene_biotype \
-G NMR_ref/NMR_ref_female.fa \
liver_hisat2/out_female/*.bam
sbatch featCounts.sh
cat *.out
--------------file download
scp -r tof4003@aphrodite.med.cornell.edu:/athena/angsd/scratch/tof4003/project/featureCounts "/Users/phoebefei/Desktop/WCM/W2022/CMPB 5004/Project/ANGSD Project tof4003"
Output showing feature count .out for 1 of the run
========== _____ _ _ ____ _____ ______ _____
===== / ____| | | | _ \| __ \| ____| /\ | __ \
===== | (___ | | | | |_) | |__) | |__ / \ | | | |
==== \___ \| | | | _ <| _ /| __| / /\ \ | | | |
==== ____) | |__| | |_) | | \ \| |____ / ____ \| |__| |
========== |_____/ \____/|____/|_| \_\______/_/ \_\_____/
v1.6.2
//========================== featureCounts setting ===========================\\
|| ||
|| Input files : 11 BAM files ||
|| S SRR5517432.f_breeder.hisat2.sorted.bam ||
|| S SRR5517435.f_breeder.hisat2.sorted.bam ||
|| S SRR5517436.f_breeder.hisat2.sorted.bam ||
|| S SRR5517437.f_breeder.hisat2.sorted.bam ||
|| S SRR5517439.f_breeder.hisat2.sorted.bam ||
|| S SRR5517440.f_breeder.hisat2.sorted.bam ||
|| S SRR5517441.f_non-breeder.hisat2.sorted.bam ||
|| S SRR5517445.f_non-breeder.hisat2.sorted.bam ||
|| S SRR5517446.f_non-breeder.hisat2.sorted.bam ||
|| S SRR5517449.f_non-breeder.hisat2.sorted.bam ||
|| S SRR5517450.f_non-breeder.hisat2.sorted.bam ||
|| ||
|| Output file : hisat2_featCounts_femaleNMR.txt ||
|| Summary : hisat2_featCounts_femaleNMR.txt.summary ||
|| Annotation : femaleNMR.gtf (GTF) ||
|| Dir for temp files : featureCounts ||
|| ||
|| Threads : 1 ||
|| Level : meta-feature level ||
|| Paired-end : no ||
|| Multimapping reads : not counted ||
|| Multi-overlapping reads : not counted ||
|| Min overlapping bases : 1 ||
|| ||
\\===================== http://subread.sourceforge.net/ ======================//
//================================= Running ==================================\\
|| ||
|| Load annotation file femaleNMR.gtf ... ||
|| Features : 328743 ||
|| Meta-features : 31806 ||
|| Chromosomes/contigs : 460 ||
|| ||
|| Loading FASTA contigs : NMR_ref/NMR_ref_female.fa ||
|| 4229 contigs were loaded ||
|| ||
|| Process BAM file SRR5517432.f_breeder.hisat2.sorted.bam... ||
|| Single-end reads are included. ||
|| Assign reads to features... ||
|| Total reads : 24624992 ||
|| Successfully assigned reads : 15990487 (64.9%) ||
|| Running time : 0.39 minutes ||
|| ||
|| Process BAM file SRR5517435.f_breeder.hisat2.sorted.bam... ||
|| Single-end reads are included. ||
|| Assign reads to features... ||
|| Total reads : 26998998 ||
|| Successfully assigned reads : 16586190 (61.4%) ||
|| Running time : 0.42 minutes ||
|| ||
|| Process BAM file SRR5517436.f_breeder.hisat2.sorted.bam... ||
|| Single-end reads are included. ||
|| Assign reads to features... ||
|| Total reads : 24683617 ||
|| Successfully assigned reads : 15471399 (62.7%) ||
|| Running time : 0.38 minutes ||
|| ||
|| Process BAM file SRR5517437.f_breeder.hisat2.sorted.bam... ||
|| Single-end reads are included. ||
|| Assign reads to features... ||
|| Total reads : 23891994 ||
|| Successfully assigned reads : 14360545 (60.1%) ||
|| Running time : 0.37 minutes ||
|| ||
|| Process BAM file SRR5517439.f_breeder.hisat2.sorted.bam... ||
|| Single-end reads are included. ||
|| Assign reads to features... ||
|| Total reads : 26804951 ||
|| Successfully assigned reads : 18015831 (67.2%) ||
|| Running time : 0.42 minutes ||
|| ||
|| Process BAM file SRR5517440.f_breeder.hisat2.sorted.bam... ||
|| Single-end reads are included. ||
|| Assign reads to features... ||
|| Total reads : 24055201 ||
|| Successfully assigned reads : 13953843 (58.0%) ||
|| Running time : 0.37 minutes ||
|| ||
|| Process BAM file SRR5517441.f_non-breeder.hisat2.sorted.bam... ||
|| Single-end reads are included. ||
|| Assign reads to features... ||
|| Total reads : 26251990 ||
|| Successfully assigned reads : 17630134 (67.2%) ||
|| Running time : 0.41 minutes ||
|| ||
|| Process BAM file SRR5517445.f_non-breeder.hisat2.sorted.bam... ||
|| Single-end reads are included. ||
|| Assign reads to features... ||
|| Total reads : 27278979 ||
|| Successfully assigned reads : 16961068 (62.2%) ||
|| Running time : 0.43 minutes ||
|| ||
|| Process BAM file SRR5517446.f_non-breeder.hisat2.sorted.bam... ||
|| Single-end reads are included. ||
|| Assign reads to features... ||
|| Total reads : 39704440 ||
|| Successfully assigned reads : 25278167 (63.7%) ||
|| Running time : 0.62 minutes ||
|| ||
|| Process BAM file SRR5517449.f_non-breeder.hisat2.sorted.bam... ||
|| Single-end reads are included. ||
|| Assign reads to features... ||
|| Total reads : 25478196 ||
|| Successfully assigned reads : 17352346 (68.1%) ||
|| Running time : 0.40 minutes ||
|| ||
|| Process BAM file SRR5517450.f_non-breeder.hisat2.sorted.bam... ||
|| Single-end reads are included. ||
|| Assign reads to features... ||
|| Total reads : 30450808 ||
|| Successfully assigned reads : 19670289 (64.6%) ||
|| Running time : 0.48 minutes ||
|| ||
|| Read assignment finished. ||
|| ||
|| Summary of counting results can be found in file "featureCounts/hisat2_fe ||
|| atCounts_femaleNMR.txt.summary" ||
|| ||
\\===================== http://subread.sourceforge.net/ ======================//Analyze Count Data
1. Import and process the summary Table
library(ggplot2)
library(magrittr)
direct <- '/Users/phoebefei/Desktop/WCM/W2022/CMPB 5004/Project/ANGSD Project tof4003/'
#hisat2 males & females
rchisat2f <-paste0(direct,"hisat2_featCounts_femaleNMR.txt.summary") %>%
read.table(., header = TRUE)
rchisat2m <-paste0(direct,"hisat2_featCounts_maleNMR.txt.summary") %>%
read.table(., header = TRUE)
#star males & females
rcstarf <-paste0(direct,"star_featCounts_femaleNMR.txt.summary") %>%
read.table(., header = TRUE)
rcstarm <-paste0(direct,"star_featCounts_maleNMR.txt.summary") %>%
read.table(., header = TRUE)a. Clean the sample names
library(stringr)
#reassign sample names
nameliststarm <- str_extract(names(rcstarm), "SRR[0-9]+.(m|f)_[a-z]+")
namelisthisat2m <- str_extract(names(rchisat2m), "SRR[0-9]+.(m|f)_[a-z]+")
nameliststarf <- str_extract(names(rcstarf), "SRR[0-9]+.(m|f)_[a-z]+")
namelisthisat2f <- str_extract(names(rchisat2f), "SRR[0-9]+.(m|f)_[a-z]+")#don't want the status to be modified
names(rcstarm)[2:length(rcstarm)] <- nameliststarm[2:length(nameliststarm)]
names(rcstarf)[2:length(rcstarf)] <- nameliststarf[2:length(nameliststarf)]
names(rchisat2m)[2:length(rchisat2m)] <- namelisthisat2m[2:length(namelisthisat2m)]
names(rchisat2f)[2:length(rchisat2f)] <- namelisthisat2f[2:length(namelisthisat2f)]b.Plotting the .summary results
library(patchwork)
library(reshape2)
#STAR M
melt_starm <- melt(rcstarm, id.vars = "Status")
names(melt_starm)[3] <- "Counts"
names(melt_starm)[2] <- "Sample"
#remove 0 instances
melt_starm <- subset(melt_starm, Counts!=0)
#add a row saying identifying this data
melt_starm["Type"] <- rep("Feature Count on Male using STAR",nrow(melt_starm))
#STAR F
melt_starf <- melt(rcstarf, id.vars = "Status")
names(melt_starf)[3] <- "Counts"
names(melt_starf)[2] <- "Sample"
melt_starf <- subset(melt_starf, Counts!=0)
melt_starf["Type"] <- rep("Feature Count on Female using STAR",nrow(melt_starf))
#Hisat2 M
melt_hisat2m <- melt(rchisat2m, id.vars = "Status")
names(melt_hisat2m)[3] <- "Counts"
names(melt_hisat2m)[2] <- "Sample"
melt_hisat2m <- subset(melt_hisat2m, Counts!=0)
melt_hisat2m["Type"] <- rep("Feature Count on Male using Hisat2",nrow(melt_hisat2m))
#Hisat2 F
melt_hisat2f <- melt(rchisat2f, id.vars = "Status")
names(melt_hisat2f)[3] <- "Counts"
names(melt_hisat2f)[2] <- "Sample"
melt_hisat2f <- subset(melt_hisat2f, Counts!=0)
melt_hisat2f["Type"] <- rep("Feature Count on Male using Hisat2",nrow(melt_hisat2f))Male Alignment : STAR vs. Hisat2
melt_m <- rbind(melt_starm,melt_hisat2m)
melt_f <- rbind(melt_starf,melt_hisat2f)#set texts size
theme_set(theme(legend.position = "bottom",legend.key.size = unit(0.4, 'cm'),axis.text=element_text(size=10), axis.title=element_text(size=10)))
#bar plot the combined graph
plotm <- ggplot(data = melt_m,aes(x = Counts, y = Sample))+geom_bar(stat = 'identity', aes(fill = Status),position = "dodge")+facet_wrap(Type~., nrow =2, strip.position = "right")+ggtitle("Plot Based on Summary .txt Files for Males")+guides(fill=guide_legend(nrow=2,byrow=TRUE))ggsave("Male Sample Feature Count Summary.png")#set texts size
theme_set(theme(legend.position = "bottom",legend.key.size = unit(0.4, 'cm'),axis.text=element_text(size=10), axis.title=element_text(size=10)))
#bar plot the combined graph
plotf <- ggplot(data = melt_f,aes(x = Counts, y = Sample))+geom_bar(stat = 'identity', aes(fill = Status),width = 1,position = "dodge")+facet_wrap(Type~., nrow =2, strip.position = "right")+ggtitle("Plot Based on Summary .txt Files")+guides(fill=guide_legend(nrow=2,byrow=TRUE))ggsave("Female Sample Feature Count Summary.png")2. Create DESeq2 Objects
a. Import readcounts and assign sample names
rchisat_f <-paste0(direct,"hisat2_featCounts_femaleNMR.txt") %>% read.table(., header = TRUE)
rchisat_m <- paste0(direct,"hisat2_featCounts_maleNMR.txt") %>% read.table(., header = TRUE)
rcstar_f <-paste0(direct,"star_featCounts_femaleNMR.txt") %>% read.table(., header = TRUE)
rcstar_m <- paste0(direct,"star_featCounts_maleNMR.txt") %>% read.table(., header = TRUE)#don't want the status to be modified
names(rcstar_m)[9:length(rcstar_m)] <- nameliststarm[2:length(nameliststarm)]
names(rcstar_f)[9:length(rcstar_f)] <- nameliststarf[2:length(nameliststarf)]
names(rchisat_m)[9:length(rchisat_m)] <- namelisthisat2m[2:length(namelisthisat2m)]
names(rchisat_f)[9:length(rchisat_f)] <- namelisthisat2f[2:length(namelisthisat2f)]b. Merge male and female data and create rowData
#create rowname and delete non-numerical data
#Star Males
row.names(rcstar_m) <- make.names(rcstar_m$Geneid)
star_m_row <-rcstar_m[,c(7:8),drop = FALSE]
rcstar_m <- rcstar_m[ ,-c(1:8)]
# Star Females
row.names(rcstar_f) <- make.names(rcstar_f$Geneid)
star_f_row <-rcstar_f[,c(7:8),drop = FALSE]
rcstar_f <- rcstar_f[ ,-c(1:8)]
#bind star
duprows <- rownames(star_f_row) %in% rownames(star_m_row)
star_row <- rbind(star_f_row, star_m_row[!duprows,])
# Hisat2 Males
row.names(rchisat_m) <- make.names(rchisat_m$Geneid)
hisat_m_row <-rchisat_m[,c(7:8),drop = FALSE]
rchisat_m <- rchisat_m[ ,-c(1:8)]
#Hisat2 Females
row.names(rchisat_f) <- make.names(rchisat_f$Geneid)
hisat_f_row <-rchisat_f[,c(7:8),drop = FALSE]
rchisat_f <- rchisat_f[ ,-c(1:8)]
#bind hisat2
duprows <- rownames(hisat_m_row) %in% rownames(hisat_f_row)
hisat2_row <- rbind(hisat_m_row, hisat_f_row[!duprows,])#merge & only keep same gene name
rcstar <- merge(rcstar_m, rcstar_f,by=0, all = TRUE)
rcstar[is.na(rcstar)] <- 0
rchisat <- merge(rchisat_f, rchisat_m,by=0, all = TRUE)
rchisat[is.na(rchisat)] <- 0
row.names(rcstar) <- make.names(rcstar$Row.names)
rcstar <- rcstar[, -c(1)]
row.names(rchisat) <- make.names(rchisat$Row.names)
rchisat <- rchisat[, -c(1)]c. Create sex and breed condition
star_level <- gsub("SRR[0-9]+.","",names(rcstar))
star_sex <- gsub("_.+","",star_level)
star_breed <- gsub("(m|f)_","",star_level)
hisat_level <- gsub("SRR[0-9]+.","",names(rchisat))
hisat_sex <- gsub("_.+","",hisat_level)
hisat_breed <- gsub("(m|f)_","",hisat_level)sampleinfo_hisat <- data.frame(breed = hisat_breed, sex = hisat_sex, row.names = names(rchisat))
sampleinfo_star <- data.frame(breed = star_breed, sex = star_sex ,row.names = names(rcstar))d. Create DESeq2 objects including the processed information
Packages for DESeq and analysis
BiocManager::install("DESeq2")
BiocManager::install("vsn")
BiocManager::install("plyranges")
BiocManager::install("apeglm")
BiocManager::install("EnhancedVolcano")
BiocManager::install("org.Sc.sgd.db") ## Loading required package: S4Vectors## Loading required package: stats4## Loading required package: BiocGenerics##
## Attaching package: 'BiocGenerics'## The following objects are masked from 'package:stats':
##
## IQR, mad, sd, var, xtabs## The following objects are masked from 'package:base':
##
## anyDuplicated, append, as.data.frame, basename, cbind, colnames,
## dirname, do.call, duplicated, eval, evalq, Filter, Find, get, grep,
## grepl, intersect, is.unsorted, lapply, Map, mapply, match, mget,
## order, paste, pmax, pmax.int, pmin, pmin.int, Position, rank,
## rbind, Reduce, rownames, sapply, setdiff, sort, table, tapply,
## union, unique, unsplit, which.max, which.min##
## Attaching package: 'S4Vectors'## The following objects are masked from 'package:base':
##
## expand.grid, I, unname## Loading required package: IRanges## Loading required package: GenomicRanges## Loading required package: GenomeInfoDb## Loading required package: SummarizedExperiment## Loading required package: MatrixGenerics## Loading required package: matrixStats##
## Attaching package: 'MatrixGenerics'## The following objects are masked from 'package:matrixStats':
##
## colAlls, colAnyNAs, colAnys, colAvgsPerRowSet, colCollapse,
## colCounts, colCummaxs, colCummins, colCumprods, colCumsums,
## colDiffs, colIQRDiffs, colIQRs, colLogSumExps, colMadDiffs,
## colMads, colMaxs, colMeans2, colMedians, colMins, colOrderStats,
## colProds, colQuantiles, colRanges, colRanks, colSdDiffs, colSds,
## colSums2, colTabulates, colVarDiffs, colVars, colWeightedMads,
## colWeightedMeans, colWeightedMedians, colWeightedSds,
## colWeightedVars, rowAlls, rowAnyNAs, rowAnys, rowAvgsPerColSet,
## rowCollapse, rowCounts, rowCummaxs, rowCummins, rowCumprods,
## rowCumsums, rowDiffs, rowIQRDiffs, rowIQRs, rowLogSumExps,
## rowMadDiffs, rowMads, rowMaxs, rowMeans2, rowMedians, rowMins,
## rowOrderStats, rowProds, rowQuantiles, rowRanges, rowRanks,
## rowSdDiffs, rowSds, rowSums2, rowTabulates, rowVarDiffs, rowVars,
## rowWeightedMads, rowWeightedMeans, rowWeightedMedians,
## rowWeightedSds, rowWeightedVars## Loading required package: Biobase## Welcome to Bioconductor
##
## Vignettes contain introductory material; view with
## 'browseVignettes()'. To cite Bioconductor, see
## 'citation("Biobase")', and for packages 'citation("pkgname")'.##
## Attaching package: 'Biobase'## The following object is masked from 'package:MatrixGenerics':
##
## rowMedians## The following objects are masked from 'package:matrixStats':
##
## anyMissing, rowMedians## Warning in DESeqDataSet(se, design = design, ignoreRank): some variables in
## design formula are characters, converting to factors
## Warning in DESeqDataSet(se, design = design, ignoreRank): some variables in
## design formula are characters, converting to factors3. Differential Expression Analysis on STAR Data
a. Access Read Counts
Cleaning the data
png("Count Information STAR.png")
par(mar=c(11, 4.1, 4.1, 2.1))
colSums(counts(DESeq.star)) %>% barplot(las=2, main = "Count Information STAR")
dev.off()dim(DESeq.star)## [1] 31806 23keep_star <- rowSums(counts(DESeq.star)) > 0
DESeq.star <- DESeq.star[keep_star,]
dim(DESeq.star)## [1] 23615 23b. Calculate Size Factors
# define a function to calculate the geometric mean
gm_mean <- function(x, na.rm=TRUE){ exp(sum(log(x[x > 0]), na.rm=na.rm) / length(x)) }
star_pseudo_ref <- counts(DESeq.star) %>% apply(., 1, gm_mean)
star_ref_ratio <- counts(DESeq.star) %>% apply(., 2, function(cts){ cts/star_pseudo_ref})
counts(DESeq.star)[1,]/star_pseudo_ref[1]## SRR5517433.m_breeder SRR5517434.m_breeder SRR5517438.m_breeder
## 1.9406255 0.0000000 0.0000000
## SRR5517442.m_non SRR5517443.m_breeder SRR5517444.m_non
## 0.0000000 0.0000000 0.0000000
## SRR5517447.m_breeder SRR5517448.m_non SRR5517451.m_non
## 0.0000000 0.0000000 0.9703128
## SRR5517452.m_non SRR5517453.m_non SRR5517454.m_breeder
## 0.9703128 0.0000000 0.0000000
## SRR5517432.f_breeder SRR5517435.f_breeder SRR5517436.f_breeder
## 0.0000000 0.0000000 0.0000000
## SRR5517437.f_breeder SRR5517439.f_breeder SRR5517440.f_breeder
## 0.0000000 0.0000000 0.0000000
## SRR5517441.f_non SRR5517445.f_non SRR5517446.f_non
## 0.0000000 0.0000000 0.0000000
## SRR5517449.f_non SRR5517450.f_non
## 0.0000000 0.0000000star_ref_ratio[1,]## SRR5517433.m_breeder SRR5517434.m_breeder SRR5517438.m_breeder
## 1.9406255 0.0000000 0.0000000
## SRR5517442.m_non SRR5517443.m_breeder SRR5517444.m_non
## 0.0000000 0.0000000 0.0000000
## SRR5517447.m_breeder SRR5517448.m_non SRR5517451.m_non
## 0.0000000 0.0000000 0.9703128
## SRR5517452.m_non SRR5517453.m_non SRR5517454.m_breeder
## 0.9703128 0.0000000 0.0000000
## SRR5517432.f_breeder SRR5517435.f_breeder SRR5517436.f_breeder
## 0.0000000 0.0000000 0.0000000
## SRR5517437.f_breeder SRR5517439.f_breeder SRR5517440.f_breeder
## 0.0000000 0.0000000 0.0000000
## SRR5517441.f_non SRR5517445.f_non SRR5517446.f_non
## 0.0000000 0.0000000 0.0000000
## SRR5517449.f_non SRR5517450.f_non
## 0.0000000 0.0000000apply(star_ref_ratio,2,median)## SRR5517433.m_breeder SRR5517434.m_breeder SRR5517438.m_breeder
## 1.0078747 0.8146126 1.0000000
## SRR5517442.m_non SRR5517443.m_breeder SRR5517444.m_non
## 1.1603763 0.8713709 1.0617461
## SRR5517447.m_breeder SRR5517448.m_non SRR5517451.m_non
## 0.7585455 0.7682993 0.9595144
## SRR5517452.m_non SRR5517453.m_non SRR5517454.m_breeder
## 0.8086858 0.8179708 0.9520958
## SRR5517432.f_breeder SRR5517435.f_breeder SRR5517436.f_breeder
## 0.9269356 0.8883713 0.8475164
## SRR5517437.f_breeder SRR5517439.f_breeder SRR5517440.f_breeder
## 0.8477974 0.8737795 0.6697203
## SRR5517441.f_non SRR5517445.f_non SRR5517446.f_non
## 0.8849278 0.9036510 1.2750958
## SRR5517449.f_non SRR5517450.f_non
## 0.8804535 1.0476977DESeq.star <- estimateSizeFactors(DESeq.star)png("Library Size vs. Size Factors STAR.png")
plot(sizeFactors(DESeq.star), colSums(counts(DESeq.star)), ylab = "library sizes", xlab = "size factors", cex = 1, main = "Library Size vs. Size Factors \n (STAR)")
dev.off()c. Normalization and Plotting Counts
Normalization
counts.star_norm <- counts(DESeq.star, normalized = TRUE)png("Boxplot of Normalized Read Count STAR.png")
par(mfrow=c(1,2))
par(mar=c(11, 4.1, 4.1, 2.1))
boxplot(counts.star_norm, main = "SF normalized \nSTAR", las = 2, cex = .6)
boxplot(counts(DESeq.star), main = "Read counts only \nSTAR", las = 2, cex = .6)
dev.off()Boxplot of log2 transformation
png("Boxplot of Normalized Logged Read Count STAR.png")
par(mfrow=c(1,2))
par(mar=c(11, 4.1, 4.1, 2.1))
boxplot(log2(counts(DESeq.star)+1), notch=FALSE, main = "Non-normalized read counts", ylab="log2(read counts)", cex = .6, las = 2)
boxplot(log2(counts(DESeq.star, normalize= TRUE)+1), notch=FALSE, main = "Size-factor-normalized read counts", ylab="log2(read counts)", cex = .6, las = 2)
dev.off()Add the normalization datasets to the object
assay(DESeq.star, "log.counts") <- log2(counts(DESeq.star, normalized = FALSE) + 1)
assay(DESeq.star, "log.norm.counts") <- log2(counts(DESeq.star, normalized = TRUE) + 1)d. Mean SD plot
library(vsn)par(mfrow=c(1,1))
par(mar=c(11, 4.1, 4.1, 2.1))
star_msd <- vsn::meanSdPlot(DESeq.star@assays@data@listData$log.norm.counts, ranks=FALSE, plot = FALSE)
star_msd_gg <- star_msd$gg + ggtitle("Sequencing depth normalized log2(read counts) for STAR") + ylab("standard deviation")
ggsave("Mean SD Plot (STAR).png")e. PCA Plot
DESeq.star.trans <-DESeqTransform(DESeq.star)pca_sex_star <- plotPCA(DESeq.star.trans,
intgroup = 'sex',
returnData = FALSE) + ggtitle("PCA Plot Based on Sex, STAR")
ggsave("PCA Based on Sex STAR.png")
pca_breed_star <- plotPCA(DESeq.star.trans,
intgroup = 'breed',
returnData = FALSE) + ggtitle("PCA Plot Based on Breed Status, STAR")
ggsave("PCA Based on Breed Status STAR.png")f. Reducing dependencies
DESeq.star.rlog <- rlog(DESeq.star, blind = TRUE)png("SRR5517434 vs SRR5517433 STAR.png")
par(mfrow=c(1,2))
plot(DESeq.star@assays@data@listData$log.norm.counts[,1:2], cex=.1,main = "Size Factor and Log2-transformed \n STAR")
plot(assay(DESeq.star.rlog)[,1], assay(DESeq.star.rlog)[,2], cex=.1, main = "Rlog Transformed \n STAR", xlab = colnames(assay(DESeq.star.rlog[,1])), ylab = colnames(assay(DESeq.star.rlog[,2])) )
dev.off()Mean SD plot after reducing dependencies
star_msd_rlog <- vsn::meanSdPlot(assay(DESeq.star.rlog), ranks=FALSE, plot = FALSE)
star_msd_rlog_gg<- star_msd_rlog$gg + ggtitle("Mean SD plot Following rlog Transformation") +
coord_cartesian(ylim = c(0,3))
ggsave("Mean SD plot Following rlog Transformation STAR.png")g. Interpret the Expression Results
Re-level the conditions
DESeq.star$sex <- relevel(DESeq.star$sex, ref = "f")
DESeq.star$breed <- relevel(DESeq.star$breed, ref = "breeder")
DESeq.star$sex## [1] m m m m m m m m m m m m f f f f f f f f f f f
## Levels: f mDESeq.star$breed## [1] breeder breeder breeder non breeder non breeder non non
## [10] non non breeder breeder breeder breeder breeder breeder breeder
## [19] non non non non non
## Levels: breeder nonRunning the analysis
DESeq.star <- DESeq(DESeq.star)## using pre-existing size factors## estimating dispersions## gene-wise dispersion estimates## mean-dispersion relationship## final dispersion estimates## fitting model and testingDESeq.star## class: DESeqDataSet
## dim: 23615 23
## metadata(1): version
## assays(6): counts log.counts ... H cooks
## rownames(23615): ENSHGLG00000000001 ENSHGLG00000000002 ...
## ENSHGLG00000031857 ENSHGLG00000031858
## rowData names(28): gene_name gene_biotype ... deviance maxCooks
## colnames(23): SRR5517433.m_breeder SRR5517434.m_breeder ...
## SRR5517449.f_non SRR5517450.f_non
## colData names(3): breed sex sizeFactorHypothesis testing
### significance levels
MIN_FC=2
MIN_L2FC=log2(MIN_FC)
MAX_QVAL=0.05
### output controls
N_TOP_GENES=100Sex
library(tidyverse)## ── Attaching packages ─────────────────────────────────────── tidyverse 1.3.1 ──## ✔ tibble 3.1.6 ✔ purrr 0.3.4
## ✔ tidyr 1.2.0 ✔ dplyr 1.0.8
## ✔ readr 2.1.2 ✔ forcats 0.5.1## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::collapse() masks IRanges::collapse()
## ✖ dplyr::combine() masks Biobase::combine(), BiocGenerics::combine()
## ✖ dplyr::count() masks matrixStats::count()
## ✖ dplyr::desc() masks IRanges::desc()
## ✖ tidyr::expand() masks S4Vectors::expand()
## ✖ tidyr::extract() masks magrittr::extract()
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::first() masks S4Vectors::first()
## ✖ dplyr::lag() masks stats::lag()
## ✖ BiocGenerics::Position() masks ggplot2::Position(), base::Position()
## ✖ purrr::reduce() masks GenomicRanges::reduce(), IRanges::reduce()
## ✖ dplyr::rename() masks S4Vectors::rename()
## ✖ purrr::set_names() masks magrittr::set_names()
## ✖ dplyr::slice() masks IRanges::slice()FOI = "sex"
Level_0 = "m"
Level_1 = "f"
star.results.sex <- results(DESeq.star, saveCols = c("gene_biotype", "gene_name"),independentFiltering = TRUE, contrast = c(FOI, Level_0, Level_1)) %>%
as_tibble(rownames = "gene") %>%
dplyr::select(gene_biotype, gene_name, gene, everything()) %>%
arrange(padj)star.results.sex %>% filter(log2FoldChange >= MIN_L2FC | log2FoldChange <= -MIN_L2FC) %>% head(n=N_TOP_GENES) %>% write.csv(.,"Results_test_sex_STAR_top100.csv")Breeder vs. non-breeder
FOI = "breed"
Level_0 = "non"
Level_1 = "breeder"
star.results.breed <- results(DESeq.star, saveCols = c("gene_biotype", "gene_name"),independentFiltering = TRUE, contrast = c(FOI, Level_0, Level_1)) %>%
as_tibble(rownames = "gene") %>%
dplyr::select(gene_name,gene_biotype, gene, everything()) %>%
arrange(padj)star.results.breed %>% filter(log2FoldChange >= MIN_L2FC | log2FoldChange <= -MIN_L2FC) %>%
head(n=N_TOP_GENES) %>%
write.csv(.,"Results_test_breed_STAR_top100.csv")h. Visualizing the Expression
Adjusted p-value histogram
png("Histogram of adj.p testing sex STAR.png")
star_hist_sex <- hist(star.results.sex$padj,
col="grey", border="white", xlab="", ylab="",
main="Frequencies of adj. p-values\n(all genes) for STAR, based on sex",
cex = 0.4)
dev.off()png("Histogram of adj.p testing breed STAR.png")
hist(star.results.breed$padj,
col="grey", border="white", xlab="", ylab="",
main="frequencies of adj. p-values\n(all genes) for STAR, breeder/non-breeder comparision",
cex = 0.4)
dev.off()Sorted table of results
Sex
star.results.sex.sorted <- star.results.sex[order(star.results.sex$padj), ]
#head(star.results.sex.sorted, 10)Breed status
star.results.breed.sorted <- star.results.breed[order(star.results.breed$padj), ]
#head(star.results.breed.sorted, 10)DE of Selected Genes
Sex
png("RAB9A DE Count Sex STAR.png")
par(mfrow=c(1,2))
plotCounts(DESeq.star, gene = "ENSHGLG00000001069" , intgroup = "sex", normalized = TRUE, main = "ENSHGLG00000001069/RAB9A \n STAR")
plotCounts(DESeq.star, gene = which.max(star.results.sex$padj), intgroup = "sex",main = "Gene with max p.adj.(=least) \n on sex (STAR)")
dev.off()png("PNPLA4 DE Count Sex STAR.png")
par(mfrow=c(1,2))
plotCounts(DESeq.star, gene = "ENSHGLG00000000567" , intgroup = "sex", normalized = TRUE, main = "ENSHGLG00000000567\n-PNPLA4 STAR")
plotCounts(DESeq.star, gene = which.max(star.results.sex$padj), intgroup = "sex",main = "Gene with max p.adj.(=least) \n on sex (STAR)")
dev.off()Breed
png("MYH7 DE Count Breed STAR.png")
par(mfrow=c(1,2))
plotCounts(DESeq.star, gene = "ENSHGLG00000002052" , intgroup = "breed", normalized = TRUE, main = "ENSHGLG00000002052\n-MYH7 STAR ")
plotCounts(DESeq.star, gene = which.max(star.results.breed$padj), intgroup = "breed",main = "Gene with max p.adj.(=least) \n for breed")
dev.off()Heatmap
Sex
star.genes.sex <- subset(star.results.sex.sorted, padj < 0.05)$gene
rlog.star.sex <- DESeq.star.rlog[star.genes.sex,] %>% assay
library(pheatmap)png("Heatmap STAR sex(no scale).png")
star_heat_sex <- pheatmap(rlog.star.sex, scale="none",
show_rownames = FALSE, main = "Sex DE genes (no scaling), STAR",
color = colorRampPalette(RColorBrewer::brewer.pal(n = 7, name =
"Reds"))(100))
dev.off()png("Heatmap STAR sex (scaled).png")
pheatmap(rlog.star.sex, scale="row",
show_rownames = FALSE, main = "Sex DE genes (row-based z-score), STAR")
dev.off()Breed Status
star.genes.breed <- subset(star.results.breed.sorted, padj < 0.05 )$gene
rlog.star.breed <- DESeq.star.rlog[star.genes.breed,] %>% assaypng("Heatmap STAR breed (no scale).png")
pheatmap(rlog.star.breed, scale="none",
show_rownames = FALSE, main = "Breed Status DE genes (no scaling), STAR",
color = colorRampPalette(RColorBrewer::brewer.pal(n = 7, name =
"Reds"))(100))
dev.off()
png("Heatmap STAR breed (scaled).png")
pheatmap(rlog.star.breed, scale="row",
show_rownames = FALSE, main = "Breed Status DE genes (row-based z-score), STAR")
dev.off()Plotting the FC Values/Shrinked FC Values
plotMA
resultsNames(DESeq.star)## [1] "Intercept" "breed_non_vs_breeder" "sex_m_vs_f"star.results.srk.sex <- lfcShrink(DESeq.star, coef = 3, type = "apeglm")## using 'apeglm' for LFC shrinkage. If used in published research, please cite:
## Zhu, A., Ibrahim, J.G., Love, M.I. (2018) Heavy-tailed prior distributions for
## sequence count data: removing the noise and preserving large differences.
## Bioinformatics. https://doi.org/10.1093/bioinformatics/bty895png("MA Plot STAR Sex.png")
par(mfrow = c(1,2))
star.results.sex$sig <- star.results.sex$padj < 0.05
plotMA(star.results.sex[,c("baseMean","log2FoldChange","sig")], alpha = 0.05,
main = "no shrinkage \n (for sex comparison), STAR", ylim = c(-4,4))
plotMA(star.results.srk.sex,
alpha = 0.05,
main = "with logFC shrinkage, STAR", ylim = c(-4,4))
dev.off()star.results.srk.breed <- lfcShrink(DESeq.star, coef = 2, type = "apeglm")## using 'apeglm' for LFC shrinkage. If used in published research, please cite:
## Zhu, A., Ibrahim, J.G., Love, M.I. (2018) Heavy-tailed prior distributions for
## sequence count data: removing the noise and preserving large differences.
## Bioinformatics. https://doi.org/10.1093/bioinformatics/bty895png("MA Plot STAR Breed.png")
par(mfrow = c(1,2))
star.results.breed$sig <- star.results.breed$padj < 0.05
plotMA(star.results.breed[,c("baseMean","log2FoldChange","sig")], alpha = 0.05,
main = "no shrinkage \n(for breed comparison), STAR", ylim = c(-4,4))
plotMA(star.results.srk.breed,
alpha = 0.05,
main = "with logFC shrinkage, STAR", ylim = c(-4,4))
dev.off()Volcano plot
library(EnhancedVolcano)## Loading required package: ggrepel## Registered S3 methods overwritten by 'ggalt':
## method from
## grid.draw.absoluteGrob ggplot2
## grobHeight.absoluteGrob ggplot2
## grobWidth.absoluteGrob ggplot2
## grobX.absoluteGrob ggplot2
## grobY.absoluteGrob ggplot2#just going to directly go with the gene name
vpstarsex <- EnhancedVolcano(star.results.sex, lab = star.results.sex$gene, x = 'log2FoldChange', y = 'padj', pCutoff = 0.05, title = "Male vs. Female, STAR")
vpstarsexsrk <- EnhancedVolcano(star.results.srk.sex, lab = rownames(star.results.srk.sex), x = 'log2FoldChange', y = 'padj', pCutoff = 0.05, title = "with logFC shrinkage, STAR")
vpstarsex + vpstarsexsrk
ggsave("Volcano Plot STAR Sex.png")vpstarb <- EnhancedVolcano(star.results.breed, lab = star.results.breed$gene, x = 'log2FoldChange', y = 'padj', pCutoff = 0.05, title = "Breeder vs. Non-breeder, STAR")
vpstarbsrk <- EnhancedVolcano(star.results.srk.breed, lab = rownames(star.results.srk.breed), x = 'log2FoldChange', y = 'padj', pCutoff = 0.05, title = "with logFC shrinkage, STAR")
vpstarb + vpstarbsrk
ggsave("Volcano Plot STAR Breed.png")4. Differential Expression Analysis on STAR Data
a. Cleaning the data
png("Count Information HISAT2.png")
par(mar=c(11, 4.1, 4.1, 2.1))
colSums(counts(DESeq.hisat)) %>% barplot(las=2, main = "Count Information HISAT2")
dev.off()dim(DESeq.hisat)## [1] 31806 23keep_hisat <- rowSums(counts(DESeq.hisat)) > 0
DESeq.hisat <- DESeq.hisat[keep_hisat,]
dim(DESeq.hisat)## [1] 23943 23b. Calculate Size Factors
DESeq.hisat <- estimateSizeFactors(DESeq.hisat)png("Library Size vs. Size Factors HISAT2.png")
plot(sizeFactors(DESeq.hisat), colSums(counts(DESeq.hisat)), ylab = "library sizes", xlab = "size factors", cex = 1, main = "Size Factor for HISAT2")
dev.off()c. Normalization and Plotting Counts
counts.hisat_norm<- counts(DESeq.hisat, normalized = TRUE)png("Boxplot of Normalized Read Count HISAT2.png")
par(mfrow=c(1,2))
par(mar=c(11, 4.1, 4.1, 2.1))
boxplot(counts.hisat_norm, main = "SF normalized (HISAT2)", las = 2, cex = .6)
boxplot(counts(DESeq.hisat), main = "Read counts only (HISAT2)", las = 2, cex = .6)
dev.off()Boxplot of log2 transformation
png("Boxplot of Normalized Logged Read Count HISAT2.png")
par(mfrow=c(1,2))
par(mar=c(11, 4.1, 4.1, 2.1))
boxplot(log2(counts(DESeq.hisat)+1), notch=FALSE, main = "Non-normalized \nread counts (HISAT2)", ylab="log2(read counts)", cex = .6, las = 2)
boxplot(log2(counts(DESeq.hisat, normalize= TRUE)+1), notch=FALSE, main = "Size-factor-normalized \nread counts (HISAT2) ", ylab="log2(read counts)", cex = .6, las = 2)
dev.off()Add the normalization datasets to the object
assay(DESeq.hisat, "log.counts") <- log2(counts(DESeq.hisat, normalized = FALSE) + 1)
assay(DESeq.hisat, "log.norm.counts") <- log2(counts(DESeq.hisat, normalized = TRUE) + 1)d. Mean SD plot
par(mfrow=c(1,1))
par(mar=c(11, 4.1, 4.1, 2.1))
hisat_msd <- vsn::meanSdPlot(DESeq.hisat@assays@data@listData$log.norm.counts, ranks=FALSE, plot = FALSE)
hisat_msd_gg <- hisat_msd$gg + ggtitle("Sequencing depth normalized log2(read counts) for HISAT2") + ylab("standard deviation")
ggsave("Mean SD Plot (HISAT2).png")e. PCA Plot
DESeq.hisat.trans <-DESeqTransform(DESeq.hisat)pca_sex_hisat <- plotPCA(DESeq.hisat.trans,
intgroup = 'sex',
returnData = FALSE) + ggtitle("PCA Plot for Sex, HISAT2")
ggsave("PCA Based on Sex HISAT2.png")
pca_breed_hisat<- plotPCA(DESeq.hisat.trans,
intgroup = 'breed',
returnData = FALSE) + ggtitle("PCA Plot for Breed, HISAT2")
ggsave("PCA Based on Breed HISAT2.png")f. Reducing Dependencies
DESeq.hisat.rlog <- rlog(DESeq.hisat, blind = TRUE)png("SRR5517434 vs SRR5517433 HISAT2.png")
par(mfrow=c(1,2))
plot(DESeq.hisat@assays@data@listData$log.norm.counts[,1:2], cex=.1,main = "size factor and log2-transformed \n(HISAT2) ")
plot(assay(DESeq.hisat.rlog)[,1], assay(DESeq.hisat.rlog)[,2], cex=.1, main = "rlog transformed \n(HISAT2)", xlab = colnames(assay(DESeq.hisat.rlog[,1])), ylab = colnames(assay(DESeq.hisat.rlog[,2])) )
dev.off()Mean SD plot after reducing dependencies
hisat_msd_rlog <- vsn::meanSdPlot(assay(DESeq.hisat.rlog), ranks=FALSE, plot = FALSE)
hisat_msd_rlog_gg<-hisat_msd_rlog$gg + ggtitle("Following rlog transformation (HISAT2)") +
coord_cartesian(ylim = c(0,3))
ggsave("Mean SD plot Following rlog Transformation HISAT2.png")g. Interpret the Expression Results
Re-level the conditions
DESeq.hisat$sex <- relevel(DESeq.hisat$sex, ref = "f")
DESeq.hisat$breed <- relevel(DESeq.hisat$breed, ref = "breeder")
DESeq.hisat$sex## [1] f f f f f f f f f f f m m m m m m m m m m m m
## Levels: f mDESeq.hisat$breed## [1] breeder breeder breeder breeder breeder breeder non non non
## [10] non non breeder breeder breeder non breeder non breeder
## [19] non non non non breeder
## Levels: breeder nonRunning the analysis
DESeq.hisat <- DESeq(DESeq.hisat)## using pre-existing size factors## estimating dispersions## gene-wise dispersion estimates## mean-dispersion relationship## final dispersion estimates## fitting model and testingDESeq.hisat## class: DESeqDataSet
## dim: 23943 23
## metadata(1): version
## assays(6): counts log.counts ... H cooks
## rownames(23943): ENSHGLG00000000001 ENSHGLG00000000002 ...
## ENSHGLG00000031857 ENSHGLG00000031858
## rowData names(28): gene_name gene_biotype ... deviance maxCooks
## colnames(23): SRR5517432.f_breeder SRR5517435.f_breeder ...
## SRR5517453.m_non SRR5517454.m_breeder
## colData names(3): breed sex sizeFactorHypothesis testing
Sex
FOI = "sex"
Level_0 = "m"
Level_1 = "f"
hisat.results.sex <- results(DESeq.hisat, saveCols = c("gene_biotype", "gene_name"),independentFiltering = TRUE, contrast = c(FOI, Level_0, Level_1)) %>%
as_tibble(rownames = "gene") %>%
dplyr::select(gene_biotype, gene_name, gene, everything()) %>%
arrange(padj)hisat.results.sex %>% filter(log2FoldChange >= MIN_L2FC | log2FoldChange <= -MIN_L2FC) %>%
head(n=N_TOP_GENES) %>%
write.csv(.,"Results_test_sex_HISAT2_top100.csv")Breeder vs. non-breeder
FOI = "breed"
Level_0 = "non"
Level_1 = "breeder"
hisat.results.breed <- results(DESeq.hisat, saveCols = c("gene_biotype", "gene_name"),independentFiltering = TRUE, contrast = c(FOI, Level_0, Level_1)) %>%
as_tibble(rownames = "gene") %>%
dplyr::select(gene_name,gene_biotype, gene, everything()) %>%
arrange(padj)hisat.results.breed %>% filter(log2FoldChange >= MIN_L2FC | log2FoldChange <= -MIN_L2FC) %>%
head(n=N_TOP_GENES) %>%
write.csv(.,"Results_test_breed_HISAT2_top100.csv")h. Visualizing the Expression
Adjusted p-value histogram
png("Histogram of adj.p testing sex HISAT2.png")
hist(hisat.results.sex$padj,
col="grey", border="white", xlab="", ylab="",
main="frequencies of adj. p-values\n(all genes) for HISAT2, sex comparision",
cex = 0.4)
dev.off()png("Histogram of adj.p testing breed HISAT2.png")
hist(hisat.results.breed$padj,
col="grey", border="white", xlab="", ylab="",
main="frequencies of adj. p-values\n(all genes) for HISAT2, breeder/non-breeder comparision",
cex = 0.4)
dev.off()Sorted table of results
Sex
hisat.results.sex.sorted <- hisat.results.sex[order(hisat.results.sex$padj), ]
#head(hisat.results.sex.sorted,10)Breed status
hisat.results.breed.sorted <- hisat.results.breed[order(hisat.results.breed$padj), ]
head(hisat.results.breed.sorted, 10)## # A tibble: 10 × 9
## gene_name gene_biotype gene baseMean log2FoldChange lfcSE stat pvalue
## <chr> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 MYH7 protein_coding ENSHGL… 25.2 -23.9 2.99 -8.00 1.25e-15
## 2 <NA> protein_coding ENSHGL… 25.9 -2.10 0.318 -6.60 4.05e-11
## 3 SERPINA6 protein_coding ENSHGL… 401. -1.45 0.244 -5.93 3.02e- 9
## 4 CCND1 protein_coding ENSHGL… 1014. 1.28 0.233 5.51 3.60e- 8
## 5 POPDC2 protein_coding ENSHGL… 17.4 -1.34 0.242 -5.54 3.10e- 8
## 6 <NA> pseudogene ENSHGL… 17.8 -1.07 0.201 -5.31 1.10e- 7
## 7 PGLYRP2 protein_coding ENSHGL… 4068. 0.864 0.170 5.09 3.63e- 7
## 8 NALCN protein_coding ENSHGL… 39.7 1.24 0.262 4.73 2.20e- 6
## 9 <NA> pseudogene ENSHGL… 10.1 2.21 0.469 4.71 2.43e- 6
## 10 <NA> protein_coding ENSHGL… 87.7 1.57 0.333 4.71 2.49e- 6
## # … with 1 more variable: padj <dbl>DE of Selected Genes
Sex
png("RAB9A DE Count Sex HISAT2.png")
par(mfrow=c(1,2))
plotCounts(DESeq.hisat, gene = "ENSHGLG00000001069" , intgroup = "sex", normalized = TRUE, main = "ENSHGLG00000001069/RAB9A\n(HISAT2)")
plotCounts(DESeq.hisat, gene = which.max(hisat.results.sex$padj), intgroup = "sex",main = "Gene with max p.adj.\n(=least) for sex, HISAT2")
dev.off()png("PNPLA4 DE Count Sex HISAT2.png")
par(mfrow=c(1,2))
plotCounts(DESeq.hisat, gene = "ENSHGLG00000000567" , intgroup = "sex", normalized = TRUE, main = "ENSHGLG00000000567\n-PNPLA4 (HISAT2)")
plotCounts(DESeq.hisat, gene = which.max(hisat.results.sex$padj), intgroup = "sex",main = "Gene with max p.adj.\n(=least) for sex, HISAT2")
dev.off()Breed
png("MYH7 DE Count Breed HISAT2.png")
par(mfrow=c(1,2))
plotCounts(DESeq.hisat, gene = "ENSHGLG00000002052" , intgroup = "breed", normalized = TRUE, main = "ENSHGLG00000002052\n-MYH7 (HISAT2)")
plotCounts(DESeq.hisat, gene = which.max(hisat.results.breed$padj), intgroup = "breed",main = "Gene with max p.adj.\n(=least) for breed, HISAT2")
dev.off()Heatmap
Sex
hisat.genes.sex <- subset(hisat.results.sex.sorted, padj < 0.05)$gene
rlog.hisat.sex <- DESeq.hisat.rlog[hisat.genes.sex,] %>% assaypng("Heatmap HISAT2 sex(no scale).png")
pheatmap(rlog.hisat.sex, scale="none",
show_rownames = FALSE, main = "Sex DE genes (no scaling)\n Aligned by HISAT2",
color = colorRampPalette(RColorBrewer::brewer.pal(n = 7, name =
"Reds"))(100))
dev.off()png("Heatmap HISAT2 sex (scaled).png")
pheatmap(rlog.hisat.sex, scale="row",
show_rownames = FALSE, main = "Sex DE genes (row-based z-score)\n Aligned by HISAT2")
dev.off()Breed Status
hisat.genes.breed <- subset(hisat.results.breed.sorted, padj < 0.05 )$gene
rlog.hisat.breed <- DESeq.hisat.rlog[hisat.genes.breed,] %>% assaypng("Heatmap HISAT2 breed (no scale).png")
pheatmap(rlog.hisat.breed, scale="none",
show_rownames = FALSE, main = "Breed Status DE genes (no scaling) \n Aligned by HISAT2",
color = colorRampPalette(RColorBrewer::brewer.pal(n = 7, name =
"Reds"))(100))
dev.off()
png("Heatmap HISAT2 breed (scaled).png")
pheatmap(rlog.hisat.breed, scale="row",
show_rownames = FALSE, main = "Breed Status DE genes (row-based z-score) \n Aligned by HISAT2")
dev.off()Plotting the FC Values/Shrinked FC Values
plotMA
resultsNames(DESeq.hisat)## [1] "Intercept" "breed_non_vs_breeder" "sex_m_vs_f"hisat.results.srk.sex <- lfcShrink(DESeq.hisat, coef = 3, type = "apeglm")## using 'apeglm' for LFC shrinkage. If used in published research, please cite:
## Zhu, A., Ibrahim, J.G., Love, M.I. (2018) Heavy-tailed prior distributions for
## sequence count data: removing the noise and preserving large differences.
## Bioinformatics. https://doi.org/10.1093/bioinformatics/bty895png("MA Plot HISAT2 Sex.png")
par(mfrow = c(1,2))
hisat.results.sex$sig <- hisat.results.sex$padj < 0.05
plotMA(hisat.results.sex[,c("baseMean","log2FoldChange","sig")], alpha = 0.05,
main = "no shrinkage \n(for sex comparison), HISAT2", ylim = c(-4,4))
plotMA(hisat.results.srk.sex,
alpha = 0.05,
main = "with logFC shrinkage\n Aligned by HISAT2", ylim = c(-4,4))
dev.off()hisat.results.srk.breed <- lfcShrink(DESeq.hisat, coef = 2, type = "apeglm")## using 'apeglm' for LFC shrinkage. If used in published research, please cite:
## Zhu, A., Ibrahim, J.G., Love, M.I. (2018) Heavy-tailed prior distributions for
## sequence count data: removing the noise and preserving large differences.
## Bioinformatics. https://doi.org/10.1093/bioinformatics/bty895hisat.results.breed$sig <- hisat.results.breed$padj < 0.05png("MA Plot HISAT2 Breed.png")
par(mfrow = c(1,2))
plotMA(hisat.results.breed[,c("baseMean","log2FoldChange","sig")], alpha = 0.05,
main = "no shrinkage \n (for breed comparison), HISAT2", ylim = c(-4,4))
plotMA(hisat.results.srk.breed,
alpha = 0.05,
main = "with logFC shrinkage\n Aligned by HISAT2", ylim = c(-4,4))Volcano plot
vpsexhisat <- EnhancedVolcano(hisat.results.sex, lab = hisat.results.sex$gene, x = 'log2FoldChange', y = 'padj', pCutoff = 0.05, title = "Male vs. Female \n Aligned by HISAT2")
vphisatsexsrk <- EnhancedVolcano(hisat.results.srk.sex, lab = rownames(hisat.results.srk.sex), x = 'log2FoldChange', y = 'padj', pCutoff = 0.05, title = "with logFC shrinkage\n Aligned by HISAT2")
vpsexhisat + vphisatsexsrk
ggsave("Volcano Plot HISAT2 Sex.png")vphisatb <- EnhancedVolcano(hisat.results.breed, lab = hisat.results.breed$gene, x = 'log2FoldChange', y = 'padj', pCutoff = 0.05, title = "Breeder vs. Non-breeder\n Aligned by HISAT2")
vpshisatsrk <- EnhancedVolcano(hisat.results.srk.breed, lab = rownames(hisat.results.srk.breed), x = 'log2FoldChange', y = 'padj', pCutoff = 0.05, title = "with logFC shrinkage\n Aligned by HISAT2")
vphisatb + vpshisatsrk
ggsave("Volcano Plot HISAT2 Breed.png")5. Clean up
zipping and saving the feature counts files
#from my home directory
cd "/Users/phoebefei/Desktop/WCM/W2022/CMPB 5004/Project/ANGSD Project tof4003"
gzip *.txtDiscussion
In this project, I have tested on the question if there are differential expression within the naked mole rat dataset, and if using different analysis tools will provide different DE results. Additionally, the Result Plot directory contains all used and supplemental figures.
Issues
There are actually 2 reference genomes, 1 for male and 1 for female, available on Ensembl for the naked mole rat – however, the gene_id convention for the males are out dated, thus I have decided to just use the female reference. The issue would be, if we are analyzing sex-related organs, for example, testis/ovaries, then the male reference genome will need to be revisited.
Top genes hit in the test for sex were actually unaccounted for specific gene names - we are missing some information in the database in general.
Future Directions
Guinea pig is a close relative to naked mole-rat, but not as related as Damara mole rat. My initial interest was to access if any of the hairlessness genes on naked mole rat is related to their longevity or cancer resistance. The Damara mole rat would be a great positive control group. However, there is no publicly available data yet that has the number of replicates to perform a DE analysis.
References
1.Bens, M., Szafranski, K., Holtze, S., Sahm, A., Groth, M., Kestler, H. A., … & Platzer, M. (2018). Naked mole-rat transcriptome signatures of socially suppressed sexual maturation and links of reproduction to aging. BMC biology, 16(1), 1-13.
2.Kim, E. B., Fang, X., Fushan, A. A., Huang, Z., Lobanov, A. V., Han, L., … & Gladyshev, V. N. (2011). Genome sequencing reveals insights into physiology and longevity of the naked mole rat. Nature, 479(7372), 223-227.
3.Fang, X., Seim, I., Huang, Z., Gerashchenko, M., Xiong, Z., Turanov, A., … & Gladyshev, V. (2014). Adaptations to a subterranean environment and longevity revealed by the analysis of mole rat genomes. Cell reports, 8(5), 1354-1364.
4.Delsuc, F., & Tilak, M. K. (2015). Naked but not hairless: the pitfalls of analyses of molecular adaptation based on few genome sequence comparisons. Genome biology and evolution, 7(3), 768-774.